Show:
The Five Big Ideas
AI, ML & Deep Learning
How Machines Learn
Types & Stages of AI
AI History & Trivia
100
This Big Idea covers how computers use sensors to extract meaning from signals, including computer vision and speech recognition.
What is Perception?
100
This subset of AI lets systems improve from data rather than from explicit programming.
What is Machine Learning?
100
This learning paradigm trains a model on labeled examples to predict an output.
What is Supervised Learning?
100
This is the only category of AI that exists today, built to excel at one specific task.
What is Narrow (Weak) AI?
100
This British mathematician proposed a famous imitation game test for machine intelligence.
Who is Alan Turing?
200
Computers construct these using data structures, then run reasoning algorithms over them to derive new information.
What are representations?
200
Deep learning gets its name from the many hidden layers in this brain-inspired structure.
What is a neural network?
200
Finding hidden structure in unlabeled data, often by grouping similar items, describes this paradigm.
What is Unsupervised Learning?
200
This earliest, memoryless AI type maps current input to output; IBM's chess machine Deep Blue is the classic example.
What is a Reactive Machine?
200
The term artificial intelligence was coined at a 1956 workshop held at this New Hampshire college.
What is Dartmouth?
300
This Big Idea is fundamentally a form of statistical inference that finds patterns in data.
What is Learning?
300
This older symbolic approach pairs a knowledge base with an inference engine to apply if-then rules.
What is an Expert System?
300
An agent, an environment, rewards, and a policy are the ingredients of this trial-and-error paradigm.
What is Reinforcement Learning?
300
This AI type uses recent observations to inform decisions, like a self-driving car tracking nearby vehicles.
What is Limited Memory?
300
This 2016 program stunned the world by beating a champion at the board game Go.
What is AlphaGo?
400
This Big Idea requires an agent to draw on culture and social conventions to infer intentions from observed behavior.
What is Natural Interaction?
400
When a model memorizes its training data and then fails on new data, it has done this.
What is overfitting?
400
Predicting a continuous number such as a house price is this supervised task, as opposed to classification.
What is regression?
400
Still hypothetical, this AI type would infer human emotions, beliefs, and intentions.
What is Theory of Mind AI?
400
The T in GPT and the 2017 paper Attention Is All You Need both refer to this neural network architecture.
What is the Transformer?
500
Under the Societal Impact idea, developing standards for fair, transparent, and accountable AI falls under this two-word umbrella.
What is AI ethics?
500
This term describes the adjustable internal values that a neural network learns during training.
What are weights (parameters)?
500
Training data skewed toward one group, producing unfair outcomes, is this much-discussed problem.
What is algorithmic bias?
500
This hypothetical intelligence would vastly exceed human ability across virtually every domain.
What is Superintelligent AI?
500
These plausible-but-false outputs from a generative model are commonly called this.
What are hallucinations?